Health Economics for Prescribers Richard Smith (MED) richard.smith@uea.ac.uk David Wright (CAP) d.j.wright@uea.ac.uk 1

Stages in economic evaluation Deciding upon study question Viewpoint taken. Alternatives appraised. Assessment of costs and benefits Identification of relevant C&B. Measurement of C&B. Valuation of C (&B). Adjustment for timing. Making a decision. Adjustment for uncertainty. Lecture 2 Lectures 3 & 4 Lecture 3 Lecture 5 93

‘Drummond’ checklist Was a well-defined question posed in answerable form? Was a comprehensive description of alternatives given? Was there evidence that effectiveness had been established? Were all the important and relevant costs and consequences for each alternative identified? Were costs and consequences measured accurately/appropriately? Were costs and consequences valued credibly? Were costs and consequences adjusted for differential timing? Was an incremental analysis performed? Was allowance made for uncertainty? Did presentation/discussion of results include all issues of concern? 93

Lecture 5: Pharmaco-economic evaluation – analysis and results Use of models Decision-tree Markov (state transition) model Uncertainty (sensitivity analysis) [checklist 9] Presentation & discussion of results [checklist 10] Indices Wider discussion – limitations, other studies, other viewpoints, generalizability, feasibility, etc Using economic evaluation QALY ‘league tables’ NICE Pharmaceutical companies and pharmacists 93

1. Use of models What is a ‘model’? Why use a model? A simplification of reality to capture the ‘essence’ of the problem with the minimum level of complexity Why use a model? To synthesise data from multiple sources To handle uncertainty & assumptions, e.g. To extrapolate from intermediate to final outcomes To predict outcomes that are unknown or are unethical to collect Handle ambiguity of clinical data and variations in interpretation 93

Types of models Descriptive Prescriptive Deterministic Stochastic describes Prescriptive suggests Deterministic certainty Stochastic probability 93

Example deterministic, prescriptive model 93

Example stochastic, prescriptive model (decision-tree) Leaves Decision node Branches Chance node 93

Analysis of decision-tree Decision tree is averaged out to get the expected value (EV) for each strategy (from decision node) EV is the sum of products of the estimates of probability of events and their outcomes (payoff) EV = 0.5x100 + 0.5x0 = £50 93

Example of decision tree for cancer screening 93

Example cont… 93

Example cont … Cost (have cancer) 0.829x£4,974=£4,124 £4,124+£1,558=£5,682 £5,682x0.004=£23 Cost (no cancer) =£24 EV(cost) screening =£23+£24=£47 93

Example cont … 93

Markov Modelling Used when disease progresses over time Patients grouped into a finite number of (Markov) states Time progresses in equal increments (Markov cycles) All events or progression are represented as transitions from one state to another with a certain probability Transitions (probability of improvement or deterioration) calculated from epidemiological and/or clinical data Spending one cycle in a given state is associated with a certain cost and a defined outcome 93

States (levels of disability) Example of Markov model States (levels of disability) A B C D E  1 Cycles (years) 2 3 4 This is a Markov model, which reflects the change in disease (levels of disability) over time (years). A could be no disability, B mild, C moderate, D severe and E death. From A a state of no disability disease could progress either to mild, moderate or severe. If it became mild then treatment could either be successful at reversing this (move back to state A) or simply slow down the progression (stay in B) or it might not work and progress to moderate disability. Ditto again, but this time if move into death then stay there.

Important points about models Structure Type of model (eg Decision-tree or Markov) Elements of model (eg nodes, branches, states) Sources of data Probability Values (cost and outcomes) Conduct of sensitivity analysis to assess impact of these on the final result 93

2. Handling uncertainty Sensitivity analysis Systematically examining the influence of uncertainties in the variables and assumptions employed on the estimated results E.g. change in a unit cost value of 10% lead to change in result of >10% (sensitive) or <10% (insensitive)? Further analysis might include Alternative (sub)perspectives Use of intermediate outcome measures Subgroup analysis 93

Importance of sensitivity analysis 93

Process of sensitivity analysis Identifying the (uncertain) variables All variables in the analysis are potential candidates Give reasons for exclusion rather than inclusion Specifying the plausible range over which they should vary Reviewing the literature Consulting expert opinion Using a specified confidence interval around the mean Recalculating results based on combinations of the best guesses, most and least conservative, usually based on… One-way analysis (each variable separately) Multi-way analysis (number of variables together) Extreme scenario analysis (all variables in extreme combinations) Threshold analysis (amount of variance needed to achieve specified result) 93

3. Presentation and discussion of results What do the decision makers want to know? Is there a health gain? Is there a cost difference? What is the relationship between cost and outcome differences? Is the cost justified by the benefit (CEA/CUA)? Is there a net gain (CBA)? Is this result robust or sensitive to parameters? 93

Overall index depends on type of economic evaluation Type of Analysis Result Consequences Costs Cost Minimisation Cost Benefit Cost Utility Cost Effectiveness Money Single or multiple effects not necessarily common. Valued as “utility” eg. QALY Different magnitude of a common measure eg., LY’s gained, blood pressure reduction. Least cost alternative. Identical in all respects. Cost per unit of consequence eg. cost per LY gained. Cost per unit of consequence eg. cost per QALY. As for CUA but valued in money. Net £ cost: benefit ratio. 93

Discussion – placing results in context Description Describe and interpret the meaning of the results Explicitly discuss any possible bias and results of sensitivity analysis Limitations Guide to interpreting and generalizing the results Need to compensate for the study’s shortcomings Relevance of the study Place the results into the decision context (see study question) Compare with other studies Within the same disease area or across interventions Macro implications What effect would the intervention have on the health of the larger population? What are the resource implications, how much would it cost to provide it for everyone under the NHS? 93

Be wary! Any conclusion that a treatment is “cost-effective” should be viewed critically Strictly true only if it dominates the comparator If the ICER is estimated, so intervention is more effective and more costly, then who is the author to say that society is willing to pay that amount for the outcome? Although can say, for example, that the intervention is cost effective when compared to other accepted interventions 93

4. Using economic evaluations QALY ‘league tables’ NICE Pharmaceutical companies and pharmacists 93

QALY League Tables Health care interventions can be compared in terms of their relative cost-effectiveness if comparable outcomes are employed (QALYs, life years) This allows analysts to place their findings in a broader context of cost-effectiveness This allows for decisions to be made about allocating resources between competing interventions (ranking of results) 93

Example league table for UK 5 ‘hip QALYs’ or 1 ‘heart QALY’? Threshold 93

Objections to their use Differences in methodology Choice of comparator Choice of discount rate Method of estimating utility values Range of costs included Application of decision rules in practice Issue of divisibility or returns to scale What ‘threshold’ should be used? 93

NICE (National Institute for Health and Clinical Effectiveness) Intended to provide “authoritative, robust and reliable guidance on current best practice” Remit to produce national guidance on Individual technologies, appraisal Management of specific conditions, clinical guidance Clinical audit Criteria for making decisions clinical priorities clinical need “broad balance of benefits and costs” guidance on resources likely to be available 93

NICE mess Does not have implementation budget Issues of transparency Lobbying by patient and provider groups Too NICE to say no? Doesn’t advise on what to disinvest in Issues of “mandatory” vs. “advisory” guidance Refusal to rank technologies in any form of hierarchy, and thus refusal to admit explicitly a cost per QALY threshold (although recently changed) but implicitly… Once we have identified what the most appropriate outcome is we need to see about measuring it. Before doing that want to highlight the general ways that this data is collected. This is similar to how we would collect cost data. As with cost data or resource use each has its merits.

Value of a QALY It appears NICE believe a QALY has a value of ~£30,000 But is this the value the public place on a QALY? Recall one of the principles of welfare economics relevant source of monetary value of health outcomes is individual WTP So, what is the public’s willingness to pay for a QALY? Current ‘cutting edge’ research in economic evaluation (team from UEA, Newcastle & Aberdeen) There are two possible outcomes which place a value on health. One is to elicit a preference for a health state which is then used to weight time in that health state. The most common way to do this is the QALY, although we will discuss others. Alternatively you can place a monetary value on health and then compare money gained to money spent. This results in a cost benefit analysis and we will discuss this in full next week.

Pharmaceutical companies and pharmacists Think about this for the next workshop when looking at checklist item 10 on “issues of concern to users”… …you will be users of this information so what issues are relevant to you? No matter what measure you use the important thing is what you do with them, use them to adjust the quantity of life in a health state to give an outcome measure which represents both quality and quantity.

Final thoughts ... Practical problems when using results of an evaluation The way in which a health care system operates The fact that there exists political rationing criteria over and above that of ‘efficiency’ – need to take account of this Economic evaluation is an aid to the decision making process – it does not make decisions It might not provide the perfect basis for decision making But the question is: is it better than anything else “Don’t let the pursuit of the perfect lead you to dismiss that which is merely very good” Alan Williams No matter what measure you use the important thing is what you do with them, use them to adjust the quantity of life in a health state to give an outcome measure which represents both quality and quantity.